1. Field of the Invention
The present invention relates to an image processing device, and, in particular, the present invention relates to an image processing device that performs image processing for correcting magnification chromatic aberration, the processing method thereof, and a program that causes a computer to execute the method.
2. Description of the Related Art
Light passing through a lens may cause magnification chromatic aberration to occur. The refractive index of the lens varies depending on the wavelength of light, and hence the magnification of an image varies depending on the color of light, thereby causing the magnification chromatic aberration to occur. For example, when a white point light source is imaged, in the captured image thereof, a state is observed in which the light source appears like extending in the radiation direction from a center position that corresponds to an optical axis, with being colored like a rainbow.
Since such a magnification chromatic aberration appears as color bleeding in the captured image, and hence become a factor for the reduction of the image quality thereof, correction may be necessary. As an example of a method for correcting the magnification chromatic aberration, material is used for a lens, the difference of refractive indexes of the material, which is due to the wavelength of light, being as small as possible with respect to the lens itself. In addition, a technique is also adopted in which the combination of lenses is optimized and an achromatic effect is utilized, thereby correcting the magnification chromatic aberration. However, in such a method in which, using the performance of the lens itself or the combinational structure thereof, correction is performed, the cost of the lens is increased. Therefore, for example, as a consumer device for which cost reduction tends to become important, the cost of a lens is increased, which is not desirable.
Therefore, image pickup devices have had functions in which magnification chromatic aberration is corrected using signal processing performed on captured-image signals. For example, in a method of the related art, with respect to an input image, the image is segmented into eight regions along a radial direction, and an edge is detected in the radial direction in each of the segmented regions. Next, a color shift amount in the detected edge portion is obtained, and furthermore, on the basis of the shift amount, a magnification difference in a different color image plane is calculated in each edge portion, for example. In addition, on the basis of a magnification difference having the highest occurrence rate, a pixel value at a position displaced from a processing-target pixel owing to the magnification chromatic aberration is obtained, and the pixel value is regarded as the pixel value of the processing-target pixel. An example of this method is disclosed in Japanese Unexamined Patent Application Publication No. 2006-20275 (FIG. 1).
In addition, in a second method of the related art, using a magnification difference calculated in each of eight segmented regions, magnification differences in four directions, namely, horizontal and vertical directions, are calculated. In addition, with respect to each of four image planes into which an image is segmented along the horizontal and vertical directions, the displacement vector of a color shift is obtained. An example of this method is disclosed in Japanese Unexamined Patent Application Publication No. 2006-20275 (FIG. 1).
In addition, as a third method of the related art, there has been adopted the following configuration for magnification chromatic aberration correction. Namely, a displacement amount is preliminarily calculated with respect to each pixel position, and furthermore, data of the displacement amount is stored, as correction data, in an image pickup device with being associated with each lens parameter. In addition, at the time of correction processing, using the correction data, a correction amount is obtained with respect to each pixel selected as a processing target. Next, on the basis of the correction amount, a pixel value at a position displaced from a processing-target pixel owing to the magnification chromatic aberration is obtained, and the pixel value is regarded as the pixel value of the processing-target pixel. An example of this method is disclosed in Japanese Unexamined Patent Application Publication No. 2006-135805 (FIG. 1).